15 October 2026
US states ask courts to block OpenAI from building self-improving AI
First reported
Platformer ran this on .
- State attorneys general are asking courts to stop OpenAI from developing AI that improves itself.
- The request says such work should only go ahead if the systems can be shown to share human values.
- The newsletter notes Republican attorneys general are backing a demand usually made by AI safety advocates.
How it was covered
PlatformerCasey Newton
The states' demand on recursive self-improvement, and the Republican split with accelerationists